Date: Wed, 12 Jun 2024 16:13:38 +0300 (EEST)
From: Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>
To: Ferry Toth <fntoth@...il.com>
cc: Jiri Slaby <jirislaby@...nel.org>, neil.armstrong@...aro.org, 
    Greg Kroah-Hartman <gregkh@...uxfoundation.org>, 
    linux-serial <linux-serial@...r.kernel.org>, 
    LKML <linux-kernel@...r.kernel.org>, Al Cooper <alcooperx@...il.com>, 
    Alexander Shiyan <shc_work@...l.ru>, 
    Alexandre Belloni <alexandre.belloni@...tlin.com>, 
    Alexandre Torgue <alexandre.torgue@...s.st.com>, 
    Alim Akhtar <alim.akhtar@...sung.com>, 
    Andrew Morton <akpm@...ux-foundation.org>, 
    "Aneesh Kumar K.V" <aneesh.kumar@...nel.org>, 
    AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>, 
    Baolin Wang <baolin.wang@...ux.alibaba.com>, 
    Baruch Siach <baruch@...s.co.il>, Bjorn Andersson <andersson@...nel.org>, 
    Claudiu Beznea <claudiu.beznea@...on.dev>, 
    "David S. Miller" <davem@...emloft.net>, 
    Fabio Estevam <festevam@...il.com>, Hammer Hsieh <hammerh0314@...il.com>, 
    Christian König <christian.koenig@....com>, 
    Christophe Leroy <christophe.leroy@...roup.eu>, 
    Chunyan Zhang <zhang.lyra@...il.com>, Jerome Brunet <jbrunet@...libre.com>, 
    Jonathan Hunter <jonathanh@...dia.com>, 
    Kevin Hilman <khilman@...libre.com>, 
    Konrad Dybcio <konrad.dybcio@...aro.org>, 
    Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>, 
    Kumaravel Thiagarajan <kumaravel.thiagarajan@...rochip.com>, 
    Laxman Dewangan <ldewangan@...dia.com>, 
    linux-arm-kernel@...ts.infradead.org, linux-arm-msm@...r.kernel.org, 
    "Maciej W. Rozycki" <macro@...am.me.uk>, 
    Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>, 
    Martin Blumenstingl <martin.blumenstingl@...glemail.com>, 
    Matthias Brugger <matthias.bgg@...il.com>, 
    Maxime Coquelin <mcoquelin.stm32@...il.com>, 
    Michael Ellerman <mpe@...erman.id.au>, Michal Simek <michal.simek@....com>, 
    "Naveen N. Rao" <naveen.n.rao@...ux.ibm.com>, 
    Nicolas Ferre <nicolas.ferre@...rochip.com>, 
    Nicholas Piggin <npiggin@...il.com>, Orson Zhai <orsonzhai@...il.com>, 
    Pali Rohár <pali@...nel.org>, 
    Patrice Chotard <patrice.chotard@...s.st.com>, 
    Peter Korsgaard <jacmet@...site.dk>, 
    Richard Genoud <richard.genoud@...il.com>, 
    Russell King <linux@...linux.org.uk>, 
    Sascha Hauer <s.hauer@...gutronix.de>, Shawn Guo <shawnguo@...nel.org>, 
    Stefani Seibold <stefani@...bold.net>, 
    Sumit Semwal <sumit.semwal@...aro.org>, 
    Taichi Sugaya <sugaya.taichi@...ionext.com>, 
    Takao Orito <orito.takao@...ionext.com>, 
    Tharun Kumar P <tharunkumar.pasumarthi@...rochip.com>, 
    Thierry Reding <thierry.reding@...il.com>, Timur Tabi <timur@...nel.org>, 
    Vineet Gupta <vgupta@...nel.org>, 
    Marek Szyprowski <m.szyprowski@...sung.com>
Subject: Re: [PATCH 00/15] tty: serial: switch from circ_buf to kfifo

On Mon, 10 Jun 2024, Ferry Toth wrote:
> Op 07-06-2024 om 22:32 schreef Ferry Toth:
> > Op 22-04-2024 om 07:51 schreef Jiri Slaby:
> > > On 19. 04. 24, 17:12, Neil Armstrong wrote:
> > > > On 05/04/2024 08:08, Jiri Slaby (SUSE) wrote:
> > > > > This series switches tty serial layer to use kfifo instead of
> > > > > circ_buf.
> > > > > 
> > > > > The reasoning can be found in the switching patch in this series:
> > > > > """
> > > > > Switch from struct circ_buf to proper kfifo. kfifo provides much
> > > > > better
> > > > > API, esp. when wrap-around of the buffer needs to be taken into
> > > > > account.
> > > > > Look at pl011_dma_tx_refill() or cpm_uart_tx_pump() changes for
> > > > > example.
> > > > > 
> > > > > Kfifo API can also fill in scatter-gather DMA structures, so it easier
> > > > > for that use case too. Look at lpuart_dma_tx() for example. Note that
> > > > > not all drivers can be converted to that (like atmel_serial), they
> > > > > handle DMA specially.
> > > > > 
> > > > > Note that usb-serial uses kfifo for TX for ages.
> > > > > """

> > > Sadly, everyone had a chance to test the series:
> > >    https://lore.kernel.org/all/20240319095315.27624-1-jirislaby@kernel.org/
> > > for more than two weeks before I sent this version for inclusion. And then
> > > it took another 5 days till this series appeared in -next. But noone with
> > > this HW apparently cared enough back then. I'd wish they (you) didn't.
> > > Maybe next time, people will listen more carefully:
> > > ===
> > > This is Request for Testing as I cannot test all the changes
> > > (obviously). So please test your HW's serial properly.
> > > ===
> > > 
> > > > and should've been dropped immediately when the first regressions were
> > > > reported.
> > > 
> > > Provided the RFT was mostly ignored (anyone who tested that here, or I
> > > only wasted my time?), how exactly would dropping help me finding
> > > potential issues in the series? In the end, noone is running -next in
> > > production, so glitches are sort of expected, right? And I believe I
> > > smashed them quickly enough (despite I was sidetracked to handle the n_gsm
> > > issue). But I might be wrong, as usual.
> > 
> > I arrived at this party a bit late, sorry about that. No good excuses.
> > 
> > > So no, dropping is not helping moving forward, actions taken by e.g. Marek
> > > Szyprowski <m.szyprowski@...sung.com> do, IMNSHO.
> > 
> > Good news is I tested on Merrifield (Intel Edison) which is slow (500MHz)
> > and has a HSU that can transmit up to 3.5Mb/s. It really normally needs DMA
> > and just a single interrupt at the end of transmit and receive for which I
> > my own patches locally. The bounce buffer I was using on transmit broke due
> > to this patch, so I dropped that. Still, with the extra interrupts caused by
> > the circ buffer wrapping around it seems to work well. Too late to add my
> > Tested-by.
> > 
> > One question though: in 8250_dma.c serial8250_tx_dma() you mention "/* kfifo
> > can do more than one sg, we don't (quite yet) */".
> > 
> > I see the opportunity to use 2 sg entries to get all the data out in one dma
> > transfer, but there doesn't seem to be much documentation or examples on how
> > to do that. It seems just increasing nents to 2 would do the trick?
> 
> Nevertheless I got this to work. Very nice. Thanks for this series.
> I am seeing only 2 interrupts (x2 as each interrupt happens twice), one for
> dma complete. The 2nd, not sure but likely, uart tx done.
> In any case the whole buffer is transferred without interchar gaps.
> 
> > So, what was the reason to "don't (quite yet)"?
> 
> Before considering to send out a patch for this, are there any caveats that
> I'm overlooking?

Not exactly related to that quoted comment, but you should Cc the person 
who added RNZ1 DMA a year or two back (in 8250_dw.c) because it required 
writing Tx length into some custom register. I don't know the meaning of 
that HW specific register so it would be good to get confirmation the HW 
is okay if it gets more than 1 sg entry (at worst, a HW-specific limit 
on nents might need to be imposed).

-- 
 i.

Powered by blists - more mailing lists